not that anyone cares:
Athlon XP 256KB- 84 mm^2
Athlon XP 512KB- 101 mm^2
Athlon 64- 193 mm^2
Northwood- 136 mm^2
Prescott- 112 mm^2
holy shit
-
- Tenth Dan Procrastinator
- Posts: 4891
- Joined: Fri Jul 18, 2003 3:09 am
- Location: San Jose, CA
Another thing to note is that doubling the number of bits tends to more than double the number of gates and therefore area needed to implement the functional units. Also, even the number of registers in the register file on the Athlon64 is doubled, so think quadruple area since there are 4x the number of bits now!
-
- Grand Pooh-Bah
- Posts: 6722
- Joined: Tue Sep 19, 2006 8:45 pm
- Location: Portland, OR
- Contact:
Re: holy shit
if I see Itanium numbers anywhere I'll post them.Dwindlehop wrote:not that anyone cares:
130 nm Athlon XP 256KB- 84 mm^2
130 nm Athlon XP 512KB- 101 mm^2
130 nm Athlon 64 1MB- 193 mm^2
130 nm Northwood 512KB- 136 mm^2
90 nm Prescott 1MB- 112 mm^2
All Athlon 64s, both regular and FX, have 1 MB of cache. The difference is in the package. The cheaper ones have fewer pins because they have a single channel memory controller. This doesn't change the die size, so both the cheap Athlon 64 and the Athlon 64 FX are 193 mm^2.
-
- Tenth Dan Procrastinator
- Posts: 4891
- Joined: Fri Jul 18, 2003 3:09 am
- Location: San Jose, CA
http://news.com.com/2100-1001-975472.ht ... _lede1_hed
The [Madison's] 374 square millimeter "die" size is a bit smaller than the 421 square millimeter size of Itanium 2, but still is large and therefore pricey. The larger the chip size, the fewer chips can be carved from each slice of silicon.
-
- Tenth Dan Procrastinator
- Posts: 4891
- Joined: Fri Jul 18, 2003 3:09 am
- Location: San Jose, CA
This paper happens to have a nice picture of an "experimental?" I2 on a .13um process with 6MB of cach. Interesting to note is that like 94% of the transistors on the chip are cache. It's SRAM with a processor and a bus interface 
On another note, it's kinda depressing to see how much silicon they're wasting in the processor section.

On another note, it's kinda depressing to see how much silicon they're wasting in the processor section.